Intel/Qwen2-57B-A14B-Instruct-int4-inc

Model Details

This model is an int4 model symmetric quantized with group_size 128 of Qwen/Qwen2-57B-A14B-Instruct generated by intel/auto-round, if you need AutoGPTQ format, please load the model with revision 730c79e .

INT4 CPU/CUDA Inference

## pip install auto-round (cpu needs version > 0.3.1)
from auto_round import AutoRoundConfig ##must import for auto-round format
import torch
from transformers import AutoModelForCausalLM,AutoTokenizer
quantized_model_dir = "Intel/Qwen2-57B-A14B-Instruct-int4-inc"
tokenizer = AutoTokenizer.from_pretrained(quantized_model_dir)

model = AutoModelForCausalLM.from_pretrained(
    quantized_model_dir,
    torch_dtype=torch.float16,
    device_map="auto",
    ##revision="730c79e" ## AutoGTPQ format
)
prompt = "There is a girl who likes adventure,"
messages = [
    {"role": "system", "content": "You are a helpful assistant."},
    {"role": "user", "content": prompt}
]

tokenizer = AutoTokenizer.from_pretrained(quantized_model_dir)
text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True
)
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)

generated_ids = model.generate(
    model_inputs.input_ids,
    max_new_tokens=50,  ##change this to align with the official usage
    do_sample=False  ##change this to align with the official usage
)
generated_ids = [
output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
]

response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
print(response)

##prompt = "请介绍一下阿里巴巴公司"
##阿里巴巴集团是一家中国跨国科技公司，成立于1999年，总部位于中国杭州。阿里巴巴是全球最大的零售交易平台之一，拥有淘宝、天猫、阿里云等知名业务。阿里巴巴的使命是“让天下没有难做的生意

##prompt = "9.8大还是9.11大"
##9.8和9.11都是小数，要比较它们的大小，我们只需要比较它们的小数部分。在这个情况下，9.8小于9.11。因此，答案是9.11大。

##prompt = "Once upon a time,"
##there was a land far, far away where magic and wonder filled the air. In this land, there lived a young girl named Lily who had a heart full of dreams and a spirit that could not be tamed. One day, while wandering throughy.

##prompt = "There is a girl who likes adventure,"
##That's great to hear! Adventure can be a wonderful way to explore new places, learn about different cultures, and challenge yourself physically and mentally. If the girl you're referring to is looking for adventure ideas, here are some suggestions:

##1. Travel

Evaluate the model

pip3 install lm-eval==0.4.4,auto-round

auto-round --model_name "Intel/Qwen2-57B-A14B-Instruct-int4-inc" --eval_bs 16  --eval --tasks lambada_openai,hellaswag,piqa,winogrande,truthfulqa_mc1,openbookqa,boolq,arc_easy,arc_challenge,mmlu,gsm8k,cmmlu,ceval-valid 

Metric	BF16	INT4
Avg	0.7062	0.7065
mmlu	0.7454	0.7363
cmmlu	0.8729	0.8637
ceval-valid	0.8744	0.8566
gsm8k 5 shots (strict)	0.7718	0.7559
lambada_openai	0.7442	0.7403
hellaswag	0.6517	0.6474
winogrande	0.7245	0.7309
piqa	0.8063	0.8134
truthfulqa_mc1	0.4333	0.4455
openbookqa	0.3400	0.3300
boolq	0.8829	0.8841
arc_easy	0.8035	0.8270
arc_challenge	0.5299	0.5538

Generate the model

auto-round \
--model_name  Qwen/Qwen2-57B-A14B-Instruct \
--device 0 \
--group_size 128 \
--nsamples 512 \
--bits 4 \
--iter 1000 \
--disable_eval \
--format 'auto_round' \
--output_dir "./tmp_autoround"

Ethical Considerations and Limitations

The model can produce factually incorrect output, and should not be relied on to produce factually accurate information. Because of the limitations of the pretrained model and the finetuning datasets, it is possible that this model could generate lewd, biased or otherwise offensive outputs.

Therefore, before deploying any applications of the model, developers should perform safety testing.

Caveats and Recommendations

Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model.

Here are a couple of useful links to learn more about Intel's AI software:

• Intel Neural Compressor link
• Intel Extension for Transformers link

Disclaimer

The license on this model does not constitute legal advice. We are not responsible for the actions of third parties who use this model. Please consult an attorney before using this model for commercial purposes.

Cite

@article{cheng2023optimize, title={Optimize weight rounding via signed gradient descent for the quantization of llms}, author={Cheng, Wenhua and Zhang, Weiwei and Shen, Haihao and Cai, Yiyang and He, Xin and Lv, Kaokao and Liu, Yi}, journal={arXiv preprint arXiv:2309.05516}, year={2023} }

arxiv github